First- and Second-Order Auxiliary Equations

• First-order and second-order auxiliary equations are methods that reduce complex higher-order differential and difference equations into coupled first-order or nonlinear systems for simpler analysis.
• They employ techniques like Riccati reduction, companion matrices, and nonlinear transforms to decouple systems and facilitate explicit solution construction in both continuous and discrete settings.
• The framework underpins stable numerical schemes and rigorous error analysis for PDEs and evolution equations, leading to improved computational efficiency and adaptive refinement.

First-order and second-order auxiliary equations play a central role in both the theoretical analysis and numerical treatment of differential and difference equations of order two or higher. The auxiliary equation framework enables the systematic reduction of higher-order problems to coupled first-order systems or to nonlinear lower-order equations, facilitating qualitative analysis, explicit solution construction, and the formation of stable, efficient numerical schemes in both continuous and discrete settings.

1. Auxiliary Variables and System Reduction: Conceptual Framework

The introduction of auxiliary variables is a classical mechanism for transforming higher-order equations—both ordinary differential equations (ODEs) and difference equations—into first-order systems. For a general $N$-th order linear ODE

$$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$

the canonical reduction introduces the companion vector

$$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$

yielding the first-order system $Y'(x) = A(x)\,Y(x)$, where $A(x)$ is the companion matrix. A generalization considers representations such as $y(x) = \sum_{n=1}^{N}y_n(x)$ with $N-1$ additional linear constraints to capture desired structural or physical properties (1803.03124).

For discrete problems, the procedure is analogous. The general second-order linear difference equation

$y_{n+2} + A_n\,y_{n+1} + B_n\,y_n = F_n$

can be decomposed by splitting $y_n = u_n + v_n$ with auxiliary sequences $u_n, v_n$ satisfying tailored update and propagation properties (Ayzatsky, 2017).

These reductions are essential for both theoretical spectral analysis and the construction of robust, stable numerical approximation schemes, particularly in the context of partial differential equations, time evolution problems, and wave propagation phenomena.

2. First-Order Auxiliary Equations: Riccati Reduction and Matrix Forms

The first-order auxiliary equation, typified by the Riccati reduction, transforms a second-order linear ODE into a first-order nonlinear equation for the logarithmic derivative $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$0. Substitution yields the Riccati equation:

$$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$1

(1803.03124). Once a particular solution is found, the general solution $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$2 follows via a simple quadrature:

$$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$3

In the discrete setting, analogous decomposition uses two auxiliary variables, $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$4 and $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$5, representing forward and backward modes, determined via a local linear system: $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$6 with $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$7 encoding the local dynamics. This leads to a two-dimensional first-order system: $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$8 where $$y^{(N)}(x) + a_{N-1}(x)\,y^{(N-1)}(x) + ... + a_0(x)\,y(x) = 0,$$9 and $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$0 are explicitly constructed from the coefficients $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$1 (Ayzatsky, 2017). Selection of $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$2 according to local Riccati-type relations can hence decouple or diagonalize the system, directly yielding separated modal solutions and facilitating stable, physically interpretable marching algorithms.

3. Second-Order Auxiliary Equations: Generalization for Higher Order

For third-order ODEs, the first-order Riccati-type transform is replaced by a nonlinear second-order auxiliary equation. Defining $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$3 and $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$4, their interrelation leads to a cubic, nonlinear ODE for $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$5:

$$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$6

with $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$7, capturing the dominant phase/amplitude structure of $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$8 (1803.03124). As in the Riccati case, finding a particular solution $$Y(x) = \begin{pmatrix} y(x) \ y'(x) \ \vdots \ y^{(N-1)}(x) \end{pmatrix},$$9 produces one base solution to the linear third-order ODE, with the remaining independent solutions accessible via reduction of order.

This methodology generalizes: for a linear $Y'(x) = A(x)\,Y(x)$0-th order ODE, auxiliary (potentially nonlinear) equations of order $Y'(x) = A(x)\,Y(x)$1 can encode the main propagation or amplitude characteristics, facilitating analytical and numerical exploration of the original problem's behavior.

4. Auxiliary System Approach in Partial Differential Equations

The derivation of auxiliary first-order systems is essential in the analysis of second-order elliptic and evolutionary PDEs, enabling both the formulation of well-posed variational principles and the design of efficient finite element schemes. For second-order elliptic problems:

$Y'(x) = A(x)\,Y(x)$2

the introduction of the flux variable $Y'(x) = A(x)\,Y(x)$3 yields the system: $Y'(x) = A(x)\,Y(x)$4 which can be written compactly as $Y'(x) = A(x)\,Y(x)$5 with $Y'(x) = A(x)\,Y(x)$6. The FOSLL* (first-order system LL*) approach further introduces auxiliary unknowns $Y'(x) = A(x)\,Y(x)$7 and constructs a least-squares functional

$Y'(x) = A(x)\,Y(x)$8

with

$Y'(x) = A(x)\,Y(x)$9

(Cai et al., 2014). The resulting mixed variational problem,

$A(x)$0

yields, upon elimination, the unique solution $A(x)$1 of the original second-order problem.

5. Numerical Schemes: Stability, Accuracy, and Auxiliary Variables

Auxiliary variable frameworks are instrumental in devising stable, high-accuracy numerical schemes for second-order evolution equations, especially on non-uniform or adaptive grids. For the operator-differential equation

$A(x)$2

the substitution $A(x)$3 recasts the problem as a system of first-order equations: $A(x)$4 which, in vector-operator notation, is $A(x)$5. Standard two-level implicit schemes (e.g., backward Euler, Crank–Nicolson) can be applied to this system. Upon elimination of the auxiliary variable, one obtains unconditionally stable three-level schemes for the original second-order problem: $A(x)$6 This construction preserves stability and second-order temporal accuracy even under variable step sizes, as confirmed by extensive numerical tests on bi-parabolic model equations.

6. Error Analysis and Adaptivity in Auxiliary System Frameworks

The auxiliary equation approach enables a precise characterization of discretization errors and facilitates the construction of rigorous a priori and a posteriori estimates. In the FOSLL* finite element context, the auxiliary variables enable quasi-optimal error bounds in the natural energy norm, independent of mesh size constraints: $A(x)$7 under standard regularity assumptions (Cai et al., 2014). For adaptive refinement and error control, explicit residual-based estimators are constructed from the elementwise and edgewise discrepancies in the auxiliary variables and the reconstructed primary variables, with proven reliability and efficiency bounds. Analogous strategies in the time domain are adopted in operator-difference schemes for evolutionary equations, where the energy norm and the structure of the auxiliary variables underpin unconditional stability estimates.

7. Interpretative Structural Features and Modal Analysis

Auxiliary system formulations provide additional analytic insight into the structure of solutions, invariants, and stability properties. In discrete systems, the decomposition into forward and backward propagating modes via auxiliary sequences directly exposes modal content and yields discrete analogues of continuous invariants (e.g., discrete Wronskians): $A(x)$8 which remain constant in the homogeneous case (Ayzatsky, 2017). Strategic choices of auxiliary variables diagonalize the propagator, simplifying analysis and computation, and generalizations via nonlinear auxiliary equations (e.g., cubic equations in the $A(x)$9 case) reveal the evolving complexity as the differential order increases and inform corresponding reduction-of-order techniques for solution construction (1803.03124).